Research Note: Large Load Cost Allocation

Research Note: Large Load Cost Allocation

Question

The local-cost guide needs to explain how grid spending can reach other customers without asserting that every new data center automatically raises household rates.

What the research establishes

Berkeley Lab's Electricity Rate Designs for Large Loads identifies several distinct problems: fair allocation of system costs, stranded-asset risk if a customer underuses new infrastructure, operational and resource-adequacy risk, risk sharing for new generation technologies, and the diverse service needs of very large customers.

Those problems appear in tariffs, special contracts, service agreements, interconnection studies, contribution-in-aid-of-construction terms, minimum bills, collateral, load-ramp requirements, exit fees, and demand-response obligations. The controlling record is usually a utility or regulatory document, not a developer's press release.

Berkeley Lab's June 2026 commentary on demand growth and electricity prices rejects an automatic relationship between load growth and higher prices. The result depends on existing capacity utilization, the cost of system expansion, and how costs are allocated in rate design. New sales can spread fixed costs when spare capacity exists. Expensive expansion with weak cost protection can move risk toward other customers.

The Virginia JLARC data-center report demonstrates the difference between a modeled risk and a realized bill. It analyzed multiple demand and infrastructure scenarios, local land and water effects, economic benefits, and policy choices. Its bill effects are conditional modeling, not evidence that any resident's past bill increase was caused by one data center.

Current policy examples show the available controls. The AEP Ohio large-load framework uses studies, long-term service agreements, load-ramp terms, collateral, minimum-demand charges, reimbursement, and exit provisions. Oregon's enacted HB 3546 directs its regulator to create a large-energy-use classification, allocate service costs, mitigate risk to other retail customers, and require qualifying contracts.

Community-cost map

Editorial use

The public guide should follow a hypothetical project through decisions and documents. It should acknowledge possible tax revenue, construction activity, permanent jobs, infrastructure investment, and digital-service value without treating a promised benefit as realized.

It should also preserve the most useful Field Note from E075: a manageable national share can still be a severe local constraint. A national percentage averages across regions. A site request arrives at a specific substation, watershed, permit authority, road network, tax base, and neighborhood.